Do 115 dB peaks on the LFE channel represent what you hear in a movie theater? - Page 3

Where did you dig up the 3 dB gain figure for signal addition?


The 120 dB SPL PEAK number is based on 3 channels of bass added to the LFE bass. 5 channels of bass will give you higher levels.


Pan_Law


"Pan Law (aka pan rule) is a recording and mixing principle that states that any signal of equal amplitude and phase that is played in both channels of a stereo system will increase in loudness by (up to) 6.02 dBSPL, provided there is perfect response in the loudspeaker system and perfect acoustics in the room."


Single (blue) and 2 channel (purple) signals electronic. 6 dB of gain for 2 identical signals being mixed together electronic.

Same acoustic results with my stereo speakers, 6 dB of gain.

Same acoustic results with my stacked subwoofers, 6 dB of gain.

Same results with my crossover unit, 6 dB of gain electronic.

Same results using a Y connector on my subwoofer input (R & L), 6 dB of gain.

Assuming a deep enough frequency such that there are no phase/modal issues, the SPL will be the same in both cases.

Yes, the SPL meter will read 6 dB higher output levels from the subwoofer if your subwoofer is capable enough. The 6 dB figure is based on three channels, not 5. Total SPL increase required from the subwoofer based on 5 channels is higher. Total SPL in the room will neither increase nor decrease.

When you change from Dolby BM Configuration 0 to Dolby BM Configuration 1, SPL output will shift from the main speakers to the subwoofer. You will net nothing in the room, but your subwoofer has to work much harder to get the same total SPL in the room.

Correct and this is what bosso is driving at. The net change in SPL will be zero, provided there are no negative phase/cancellation issues between the various individual sources.

No one is doubting that when two identical (SPL and Freq and phase) sources are added, the result will be +6 dB. That was really never in question (hopefully).

Naturally the 18 dB delta when five speaker channels and the LFE channel are summed at the same encoding level is constant/fixed. Say for example the bass peak was -10 dBFS (instead of 0 dBFS) in all six channels, the result would be:

dB/Pascal
105.00 3.556559
95.00 1.124683
95.00 1.124683
95.00 1.124683
95.00 1.124683
95.00 1.124683

dB Summed Value = 113.24

Read post # 31 by Bosso. He used +3 dB in his calculations instead of +6 dB in that post and in all other calculations.

Is there something in the water?

Of course I said that 105dB + 105dB = 108dB, because it does... period.

You're saying it adds to 111dB.

The source that's being fed to each channel is encoded at 0dBFS. That source at that level requires the amp/speaker combination to deliver a peak of 105dB. The summing block, the number of signals or any other argument you've presented... irrelevant.

The Delta L for 5 equal loud sound sources is 7dB, NOT 15+dB.

You have 5 sources, each delivering 105dB to your meter at the LP. They sum to 112dB (111.98 dB, more precisely).

You add 1 source that's delivering 115dB to your meter at the LP.

111.98dB + 115dB = 116.75dB.

If you have any 2 sources combining to result in any higher level than the math formula dictates, you have a flawed system that needs to be corrected by whatever means.

The example I offered is that of a calibrated sub that you then place another equally calibrated sub of equal displacement and input power on top of.

Yes, it will be 6dB louder. BUT, you then have to recalibrate the system DOWN 6dB. The result is that you now have 6dB of headroom... NOT +6dB of loudness at the LP.

Use the calculator yourself and input 5 sources at 105dB and 1 source at 115dB and tell me what the total SPL is:

http://www.sengpielaudio.com/calculator-spl.htm

If, as Ed has confirmed, there is no difference when those 5 equal sources are summed and redirected to the sub for the sub to reproduce, then what are we arguing about?

Bosso

Not sure what the discussion is about here but this caught my eye. If we're talking about two sources, each playing at 105dB the math is:

105dB = 20 Log(x) => x=177827.941
105dB = 20 Log(y) => y=177827.941
x+y = 355655.882
105dB + 105dB => 20 Log(355655.882) => 111.0205999 dB

You feeling alright Bosso??

Bosso, the math tells me I should get a 6 db gain when adding two equal signals. When I do this with my speakers, REW is showing me a 6 db gain. You might get only 3 db with random frequency random phase pink noise, but I am definitely seeing a real world, measured at my seat 6 db gain when I measure sine sweeps on 2 channels verses 1.

a bit nit picky, but it should be (x/20 micropascals) since db is a relative scale and the db SPL scale has been defined as the sound preasure level relative to 20 mpa. Since you are converting back to db, the double ommision cancels out, but that won't work if you are doing a one way conversion.

Thanks for pointing that out. So it would look like:

105dB = 20 Log(x/20) => x=3556558.82
105dB = 20 Log(y/20) => y=3556558.82
x+y = 7113117.64
105dB + 105dB => 20 Log(7113117.64/20) => 111.0205999 dB

Then please tell me how these 10 tones sum to 100dB?



If your math was correct, the 4 tones between 30-50Hz alone would be over 102dB.

Bosso

M,

Just out of interest, how are you going about this through REW? Are you running a specific freq with one and then 2 subs? Thanks.

James

You missed the boat. When someone mixes a movie soundtrack and they want high levels in the bass department, they send the same signal to multiple channels.

Your graph above is for 10 different signals being mixed together on a single channel. I don't have a clue on what that adds up to be in total SPL.



Here is pink noise. What is the SPL level of this pink noise?

I am confused. Let's say I send a 1-watt signal to one 12 inch speaker and get, say 95 dB output. If I send the same signal, but now at 2 watts, to the same speaker, I theoretically will get 98 dB output, a 3 dB gain.

If I send the 1 watt signal to 2 identical 12 inch speakers, as I understand it, I should get 98 dB output, a 3 dB gain.

If I send a 2 watt signal to the two identical 12 inch speakers, I should get a 101 dB output, right?

So if I split the signal into two identical 1 watt signals and send each to one of the 12 inch speakers, I should get 101 dB (6 dB over the 1 watt signal into 1 12 inch speaker)?

This is how it's been explained to me, unless I mistranslated what I've heard and read over time . . ..

Only use three main channels plus the LFE channel for this 120 dB SPL PEAK output calculation.


Link to Dolby on LFE


"The LFE channel carries additional bass information to supplement the bass
information in the main channels. The signal in the LFE channel is calibrated
during soundtrack production to be able to contribute 10 dB higher SPL than the
same bass signal from any one of the screen (front) channels. Even if all three
screen channels are active, enough bass could be delivered by the LFE channel
alone to bring the theatre’s subwoofer into acoustic balance with the screen
channels. This allows filmmakers to unburden the main channels by diverting the
strongest bass to the separate LFE channel, as needed. Under the most demanding
program conditions, where the bass is fully loading the left, center, and right
channels, the LFE channel could increase the bass intensity by up to 6 dB."

No, I shut the sub off (direct mode) and sent the same signal to the FR and FL one at a time and together. They have different curves because they are not symerrical in the room, but where the curves overlapped, the combined curve was 6 db higher.

You can also look at the target curves for a standard crossover. At 80 hz XO with a 75 db target, the sub target is at 69 db at 80 hz, and the speaker target is also at 69 db. The target curves are expecting 69 + 69 = 75.

That is half correct. If you double the power to the SAME driver, you get 3 db in gain. If you double the power by sending an equal signal to a second driver you get a 6 db gain. Magic? No.

In order to double the excursion on a driver, it takes four times as much power (2 squared = 4). So to double the pascals of sound energy, ie our 6 db gain, you would need 4 watts, or 1x2 = 3 db and 2x2 = another 3 db for a total of 4 watts and 101 db, or if you only double once, 2 watts and 98 db.

If you double the number of drivers, you are getting an efficiency gain. 1 watt to one driver plus one watt to the other driver will give you twice the air movement for twice the power, so 2 watts total gives a 6 db boost, or 101 db. Nice huh? Lots of drivers with a little movement will give you more output per watt than one driver will a huge XMAX.

Before this deteriorates into poo-flinging, perhaps I can get some answers to a few questions as there are some very knowledgeable posters here.

Let's say reference level for a sub is 115db on peaks. Is there a limit to the frequencies a sub is expected to reproduce at 115db? I know the upper limit is 120hz, but what about the lower limit? I don't think most movie theatres go terribly low into the bass but I'm just wondering if there's a "hard limit" specified at the lower end? I suspect not since we see some serious low bass in some soundtracks.

I believe the spec allows for a signal range from 3hz to 120hz but no one in their right mind expects a full size theater to deliver a flat response to 3 hz. I would assume there is a theater spec that has some minimum reproduction requirements in order to get a THX certification for example, but I have no clue what they are.

Just to clear up this calculator issue, we seem to be talking about apples and oranges. You are talking incoherent noise and not "5 equal sources". I am talking about the same sound being used to maximize sound levels in a mixed sound effect.


Type of sound content required in order to produce valid results from that sound level calculator.


Quote:

"Total level adding - summing up to ten incoherent noise sources

For beginners: incoherent means the signals of the two channels are irrelative like a violin and a trumpet, that means having no correlative relationship. Sometimes we say uncorrelated when we mean incoherent."

Yes, which is what you measure at the LP with your SPL meter.

I did this measurement a while back, but must have ditched it long ago, so I did it again and put reference lines on it.

The GOLD trace is the Left speaker only.
The PURPLE trace is the Left + Right speakers.
The GREEN trace is all 5 speakers.

The RED reference lines are: Base, Base + 3dB and Base + 7dB.
The BLACK reference lines are: Base(L), Base + 6dB(L+R) and Base + 14dB(L+R+SL+SR+C).



I then did a near field (well, near enough, in my haste) of the bass redirected to the sub. I couldn't graph the left speaker only, so it's L+R vs L+R+LS+RS+C.



Lo and behold, it's an increase of 8-10dB, twice what I have been saying it should be.

This led me to find the reasons for the 10.log vs 20.log numbers.




I've maintained that the 5 sats are random phase, regardless of the source, because of their placement in the room (any room) relative to each other.

My measurements verify this, and fed into my belief that the SPL calculators (I have several bookmarked and all of them calculate random phase) I've always used and all calculate dBSPL sums with the same formula, were using the correct formula.

I then placed the SPL meter in the center of the FL and FR speakers.

I shut the right speaker off and generated a 1KHz tone, which I leveled to 90dB.

I then shut the left speaker off and leveled the right speaker to 90dB.

This was so that I could be certain that each speaker separately was reading exactly the same at the meter.

I then turned both speakers on and, whamo!, 96dB.

I now see the difference between in phase vs random phase, 10.log vs 20.log, power vs pressure, SPL(pressure) vs SPL(power).

This explains the confusion with actual differences in SPL from identical co-located vs random phase sub placement in systems of multiples.

This leaves the problem of the sub outputting a sum of in phase signal vs the sats being a sum of random phase vs calibration methods, which I intend to look at more closely.

In any case,

Hats off to you, JPC, for pushing me to get to the bottom of this.

Bosso

Very interesting point... Are we boosting the LFE above the intended level by taking the content below the crossover from all of the channels, possibly intended to be random phase, and summing them in phase for the subwoofer to reproduce

If the speakers are properly Calibrated with respect to distance/delay/phase you should get the 20log 6 db type addition either way. Assuming the source calls for an identical signal in terms of frequency and phase. If the signals are not identical, you should see more like the 3 db gain, again whether they are added electonically, or acoustically at the seats. I don't see any disconnect here, other than bosso may need to tweak his distance settings on the surrounds.

In fact, in the SubMersive thread there was discussion about using special tones to fine tune your surrounds, but this has me thinking you might be able to do a "tweak until you get the full 6 db gain test" to get your sl and fl and your sr and fr inphase with each other. Hmmmmmmm.

Bingo.

I assure you, my sats are meticulously placed equidistant to the LP to the 1/16 of an inch to all 5 matched tweeters.

One question would be how exactly do you 'tweak the distance settings on the surrounds"?

Looking at the average Joe's calibration, it's done with the AVRs limited band pink noise and the RS meter, which tells you pretty much nothing, except the total SPL is matched within that limited band, when one speaker at a time is playing.

The FR changes when all 5 sats are involved, which can utterly destroy your calibration.

Even sine sweep measurements are mostly done with only the FL/FR and sub.

As I said, I'll be looking at ways to tweak the system other than the conventional method. The 'tweak to 6dB' idea is interesting.

Bosso

The discussion on the disc tones starts with Sean's post 2101 here:

http://www.avsforum.com/avs-vb/showt...7#post17040807

I would think you should be able to get tones below 120 hz in phase fairly easily. It should be doable up to 500-600 hz with the half foot steps in most receiver/processors. For higher frequencies you will need to get out the micromter and put your head in a vise, but I doubt that will improve your movie watching enjoyment

I suppose when you put it that way.....

Hey, thank goodness we get along well. Your initial posting had me doubting myself at first. Before I posted I had to go back and look at my own numbers, some of which I posted in posting 31.


My main concern was your numbers understated the SPL output that is required out of a subwoofer(s) if the end user wanted to playback at the calibrated "reference level" and used one of the recent action flicks that goose the bass output by use of that coherent audio method. LFE and redirected bass can add up to some big numbers. On a relative scale, the requirement to produce an extra 6 dB SPL of PEAK bass is a big deal.


Thank goodness most of us do not playback at full reference level. For that matter, I am not sure how many home systems are even capable of playback at full reference level.









The way that I understand it, most content in different channels is indeed incoherent. Only special bass sound effects are / may be mixed in a coherent manner.

That is why Dolby says "under the most demanding program conditions, where the bass is fully loading the left, center, and right channels, the LFE channel could increase the bass intensity by up to 6 dB."

Anyhow, I don't think that any special calibration routine is required. Standard calibration will give you the same relative levels that the sound mixer used.

Would that be measured to the left ear or right ear?








It does not matter from my experience.







I measure all adjacent speaker pairs with pink noise to observe the speaker to speaker interaction. The reason for doing that is sound pans in a sound mix are done left to right or front to back. TrueRTA is the measurement tool.

Could y'all all just go watch a movie or something like that, maybe?